作者: J. Krauth
DOI: 10.1007/978-3-642-60187-3_19
关键词:
摘要: A finite sequence of independent binomial variables is considered for which the null hypothesis homogeneity to be tested against alternative hypotheses one or two change-points, respectively. The tests are based on likelihood ratio statistic an approximate linearization log statistic. This yields scan type cusum statistics in a discrete situation. At same time maximum estimates locations change-points derived. For problem with — corresponding detection cluster we distinguish between cases fixed and variable distance points. Under exact upper bounds tails In special case Bernoulli these considerably simplified.